AI Answer Audit
"You can't measure AI visibility." Here's exactly how we measure it.
The objection is fair, and we built the method around it: ask an assistant the same question twice and the answer can move. So we don't ask once. Every question runs five times on ChatGPT, Perplexity and Google AI Overviews, and we report how often your brand was named. Every run is logged and carries an ID, so you can check the work.
- Three AI assistants, asked directly — ChatGPT, Perplexity and Google AI Overviews. We reach them through official, paid-for channels, not by copying screenshots.
- We ask every question five times on each assistant. One answer is only one result. Asking again shows whether it holds. We report how often your brand came up: ‘named 4 of the 5 times we asked’, never a yes-or-no verdict.
- We tell you how sure we are. Answers vary, so the real figure could sit about 10 points either side of the one we show. When two brands are that close, we call them level — we do not rank one above the other.
- Every AI quote is verbatim, with a run ID. The evidence pack logs every answer we collected: run ID, timestamp, engine and a fingerprint. Every quote in the report traces back to the run it came from.
- Dated, and good for about 4–8 weeks. Answers change, and we measure how much rather than hide it.
- Published method, not a black box. How we measure is on our methodology page, and the full appendix ships with the report.
"Which engines do you check — and why not Gemini?"
We measure ChatGPT, Perplexity and Google AI Overviews via official APIs and licensed SERP data. Gemini is excluded in v1 because its API terms forbid caching, analysing or "otherwise learning from" grounded results — we won't break a vendor's terms to pad the count.
"A snapshot is worthless — AI changes constantly."
We report a range, not a single number: the real figure can sit about 10 points either side of the one we show. The report is stamped with the day we measured, and is typically good for about 4–8 weeks.
"Can you guarantee AI will recommend us?"
No — and anyone who promises that is bluffing. We show you exactly where you stand and which fixes will do the most; where the engines put you is their decision.
"Isn't this just SEO?"
No. A traditional SEO audit doesn't look at what ChatGPT, Perplexity and Google AI Overviews actually say when a buyer asks — we measure exactly that. It sits alongside your SEO rather than replacing it.
What this measures: what three AI assistants said about your brand on the day we asked. Not a certification, and not a promise about where they put you next month.